国立国語研究所学術情報リポジトリ

A Reconstruction of Ancient Vietnamese

Initials Using Chu Nom Materials

著者(英) Masaaki SHIMIZU

journal or NINJAL Research Papers
publication title
number 9
page range 135-158
year 2015-07
URL [Link]
国立国語研究所論集 (NINJAL Research Papers) 9: 135–158 (2015) 135
ISSN: 2186-134X print/2186-1358 online

A Reconstruction of Ancient Vietnamese Initials Using Chữ Nôm Materials

SHIMIZU Masaaki
Osaka University / Project Collaborator, NINJAL

Abstract
In previous research concerning the Ancient Vietnamese phonology using Chữ Nôm materials,
only a few features, such as initial consonant clusters, have been considered. This study aims to
reconstruct the phonological system of Ancient Vietnamese initials, using the Sino-Vietnamese
version of the Buddhist sutra Phật Thuyết Đa ̣i Báo Phu ̣ Mẫu Ân Troṇ g Kinh 佛説大報父母恩
重經, as well as the Proto Viet-Muong phonemes established by Ferlus (2009). The Chữ Nôm
characters examined in this study allow us to consider topics such as spirantization and the voic-
ing of medial consonants—possible thanks to the existence of a number of disyllabic words con-
tained therein. Linguistic analysis of such texts does and will play an important role in filling the
missing link between Proto Viet-Muong and Middle Vietnamese (17C).*

Key words: Ancient Vietnamese, Chữ Nôm characters, Sino-Vietnamese readings

1. Preface
One of the difficulties of using Chữ Nôm (CN) characters in historical phonology comes from
the fact that they were created based on the Sino-Vietnamese (SV) reading of each Chinese
character: i.e., the Vietnamized reading of Chinese characters (Nguyễn Taì Cẩ n 1995). It is also
known that Sino-Vietnamese readings were derived from the phonological system of Middle
Chinese (MC) (Mineya 1972). This is quite different from the case of some Manyogana―the
“α group” by Mori (1991)―which were based directly on the reading of MC itself. Therefore, in
some cases such as (1a), SV readings reflect the original sound of MC, while in others, such as
(1b), they do not because of sound change and lack of identical phonemes in MC.
(1) a.
MC  SV   AV†
見 *k > k k (c/k/q)
溪 *kʰ > kʰ kʰ (kh)
曉 *h > h h (h) etc.
b.
MC SV AV
幫 *p > ɓ ɓ (b)
β ((b1)
微 *ɱ > v v (v)
(w (v)) etc.
†
Ancient Vietnamese (AV) denotes the stage of language between the formation of the SV readings and the
time of compiling Hua yi yi yu 華夷譯語 in the history of Vietnamese.
* This paper is the revised version of the oral presentation entitled “A Phonological Reconstruction of An-
cient Vietnamese Using Chữ Nôm Materials,” presented at NINJAL on Dec. 23, 2013. This study is a result
of the Core Research Project “Typological and Historical/Comparative Research on the Languages of the
Japanese Archipelago and their Environs” headed by Prof. John Whitman.
1 The form written in Dictionarium Annamiticum, Lusitanum, et Latinum of Alexandre de Rhodes (1651),
which corresponds to v in the modern orthography.
136 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015)

Those who are willing to use Chữ Nôm characters in historical phonology need to pay spe-
cial attention to the latter case. Considering this point, in this study, I would like to propose an
appropriate method of using Chữ Nôm materials in the historical phonology of Vietnamese.

2. Materials
The Chữ Nôm data used in this study are gathered from the Sino-Vietnamese version of the
Buddhist sutra Phâ ̣t Thuyết Đại Báo Phụ Mẫu Ân Trọng Kinh 佛説大報父母恩重經, which was
introduced by a number of philologists inside and outside of Vietnam. There are two different
̀ g Thi ̣Ngo ̣ 1999,
opinions on the period of this material: one regards it as from the 15th century (Hoan
Shimizu 1996), and the other regards it as from the 12th century (Nguyễn Taì Cẩn 2008, Trần
Tro ̣ng Dương 2011).
I prefer the former opinion for two reasons: (1) two taboo characters are used to avoid
the names of the kings in the Lê dynasty (15th century); and (2) the pictures inserted in the
Vietnamese version of the sutra are exactly the same as those in the Korean version edited in
1486 and not earlier.
Advocates of the latter opinion argue the quite old characteristics of Chữ Nôm contained
in this material. However, we need to be reminded that the syllable structure of Vietnamese
remained CCVC/T until as late as the 17th century (Gregerson 1969). Therefore, it is quite pos-
sible that the disyllabic structure extracted from a number of Chữ Nôm characters still existed in
the 15th century.

3. Methods
3.1 Principle
First, I will explain the essential method of reconstructing the value of each phoneme at the
stage of creating each Chữ Nôm. The data used here are (1) the phonemes of Proto Viet-Muong
(PVM) (Ferlus 2009), Middle Vietnamese (Gregerson 1969), and modern Vietnamese, and
(2) the phonemes of MC, 17th century Sino-Vietnamese (Shimizu 1999), and modern Sino-
Vietnamese. The reading of a certain Chữ Nôm in the material should be placed between PVM
and the 17th century, and the Sino-Vietnamese reading of its phonetic component is placed
between MC and the 17th century, in order to properly estimate their value at the period of creat-
ing the Chữ Nôm character. For example, there are two Chữ Nôm characters that both express
the word rắn ‘snake’ pronounced /zan5/ in the modern Northern dialect (NV) but each having a
different phonetic component with different Sino-Vietnamese readings: one is 散 tán /ta:n5/, and
the other is 吝 lâ ̣n /lʌn6/. One of the purposes here is to reconstruct the phonemic value of each
character and place it in the course of phonological change. The process of phonological change
of the initial consonants is *s > ɽ > ʐ > z for both Chữ Nôm characters, *s > t for the SV of 散,
and *l for 吝. This process can be illustrated as shown in (2) and (3):
(2) 散 rắn /zan5/ (SV: tán /ta:n5/)
AV  17C   20C   NV
SV *s- (MC) > *s- > t- > t- > t-
CN *s- (PVM) > *s- > ɽ- > ʐ- > z-
The phonemes in are those estimated for the period of creating the character.
 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015) 137

(3) 箴 rắn /zan5/ (吝 SV: lận /lʌn6/)

AV  17C   20C   NV
SV *l- (MC) > *l- > l- > l- > l-
CN *s- (PVM) > *s- > ɽ- > ʐ- > z-
We then search for the stage where the Chữ Nôm reading and the SV reading of its phonetic
component are the closest. In the case of 散, it is the stage of Ancient Vietnamese (AV), while
in 箴, it is 17C. In this way, if we follow the same process for all the phonemes, the phonological
system of AV can be reconstructed theoretically.

3.2  Application for the case of spirantization

Shimizu (1996) points out that a number of Chữ Nôm data, regarded as the traces of disyllabic
words, can be found in the present material. Some examples of these will be presented here to
apply the method above for the process of spirantization (Ferlus 1982, 2009), a phonological
process involving consonants in the medial position and an important process that occurred
between PVM and AV, which can be summarized as follows in (4):
(4) PVM Orthography   NV
a.
*p *b > v v
b.
*t *d > d z
c.
*s > r z
d. *c *ɟ > gi z
e. *tʃ > gi z
f.
*k *g > g/gh ɣ / (C)ǝ_V(C)/T
The same initial consonants occurring in the monosyllabic words went through the process
in (5):
(5) PVM Orthography NV
a. *p *b > *p > b ɓ
b. *t *d > *t > đ ɗ
c. *s > *s > t t
d. *c *ɟ > *c > ch c
e. *tʃ > *tʃ > x s
f. *k *g > *k > c/k/q k / #_V(C)/T
The cases in (6a)–(11a) demonstrate the processes of (4c) and (4a). The first example, (6a), is
a case of (4c).
(6) a. Ch: 銅 狗 鐡 蛇
CN: 主 同 破散 可列
chó đồng rắn sắt
dog copper snake iron
‘copper dog and iron snake’ (29a5)2

2 “29a5” means “page 29, first half, line 5” of the Sino-Vietnamese version of the Buddhist sutra Phật Thuyết
Đa ̣i Baó Phu ̣ Mẫu Ân Tro ̣ng Kinh 佛説大報父母恩重經.
138 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015)

The first line is the original Chinese sentence, while the second one is the Vietnamese trans-
lation with Chữ Nôm characters. The words in italics on the third line are the transcription in
the present orthography called Quốc Ngữ, followed by the glosses for each word on the fourth
line and the translation of the whole expression on the fifth line. The objective here is the word
for ‘snake,’ as we saw in the case of (2), but with the Chữ Nôm transcribing the presyllable: 破.
(6) b. PVM3 AV 17C 20C NV cf. Ru ̣c4
rắn *p-səɲʔ *pʰa-san5 ɽan5 ʐan5 zan5  pusiɲ3
破散 *pʰa:5 sa:n5 (ta:n5) (ta:n5) (ta:n5)
Referring to the PVM form and that of Ru ̣c, for instance, it is quite certain that the pres-
ent initial r comes from *s and that the SV initial of 散 also comes from the MC *s (心母).
Therefore, the most rational measure is to reconstruct both of these initials as *s at the stage of
AV. Furthermore, the SV initial of 破, which comes from the MC *pʰ (滂母), also reinforces the
reconstruction of the PVM presyllable *p-. The process of change occurring in the initial *s of the
main syllable can be summarized as shown in (7):
PVM AV  17C  20C NV
(7)
*-s- -s- > ɽ- > ʐ- z- (r)
The examples in (8) to (11) are all concerned with the process of (4a). The first case, (8a), con-
tains one character 盃司 with two components, both expressing the sounds 司 for the presyllable
and 盃 for the major syllable.
(8)
a.
Ch: 歡 愛 長 乖
CN: 弄 盃司 酉 波亘 利
lòng vui dấu hằng lìa
heart pleasant love long apart
‘The heart of pleasant love is far away for a long time.’ (20a3)
Again, referring to the PVM and Ru ̣c forms, it is certain that the initial consonant of vui had
gone through the process of spirantization. According to Ferlus (1982), the process of spirantiza-
tion of the PVM labials *p and *b can be traced as shown in (4a)’:
(4)’ a. PVM  17C 20C
*-p- > *-ɸ- *-β- > β- > v- (v)
*-b- > *-β-
Looking through the phonetic components of Chữ Nôm characters that express v in the
modern orthography, the voiced stop ɓ, such as in the SV reading of 盃 bôi /ɓo:j1/, is distin-
guished from v, as in the SV reading of 爲 vi /vi:1/ in 轗5, which is another Chữ Nôm character
that expresses the same word vui ‘pleasant.’ I believe that the labiodental feature of the latter
expresses the stage of around 20C when the process β > v had been completed. Meanwhile, the
voiced bilabial stop ɓ in ɓo:j1 can be understood in two ways. One is to regard ɓ as transcribing

3 The Proto Viet-Muong (PVM) forms are cited from Ferlus (2007).
4 One of the languages of the Chứt group belonging to the Viet-Muong branch of Mon-Khmer family.
5 Trần Văn Kiê ̣m, L.m. An-tôn (2004: 879).
 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015) 139

the stage of the voiceless stop *p of the same phoneme, as illustrated in (8b), and the other is to
regard it as expressing the stage of the voiced bilabial fricative β, as illustrated in (8c).
b. PVM AV 17C 20C NV cf. Ru ̣c
(8)
vui *t-puːj *tɨ-pu:j1 βu:j1 vu:j1 vu:j1 tapuj1
司盃 *tɨ1 ɓo:j1 (ɓo:j1) (ɓo:j1) (ɓo:j1)
c. PVM AV 17C 20C NV cf. Ru ̣c
vui *t-puːj *tɨ-βu:j1 βu:j1 vu:j1 vu:j1 tapuj1
司盃 *tɨ1 ɓo:j1 (ɓo:j1) (ɓo:j1) (ɓo:j1)
cf. 𢠿 [爲] *vi:1
The case of (8c) is more likely than the case of (8b) for two reasons. First, it is difficult
to imagine that the opposition of voiced and voiceless stops remained up to the stage of AV.
Second, it is also quite difficult to imagine that the initial stop of the SV reading of 盃 was still
voiceless or that it was useful for transcribing the voiceless stops at the time.
Example (9a) is the same as (8a). For the same reasons as above, I believe the voiced bilabial
stop in this example transcribes the voiced bilabial fricative.
(9) a. Ch: 救 父 王
CN: 助 布司 吒
trợ vua cha
save king father
‘save the father king’ (5a1)
b. PVM AV 17C 20C NV
vua - *tɨ-βuɤ1 βuə1 vuə1 vuə1
司布 *tɨ1 ɓo:5
Another problem is concerned with the height of tones. For example, the problem with (10a)
is how the tone height should be reconstructed.
(10) a. Ch: 何 不 早 死
CN: 牢 庄 阿盃 折
sao chẳng vội chết
why not early die
‘Why not die early?’ (38a1)
The modern form vội belongs to the lower tone, while the SV reading of 盃 ɓo:j1 belongs to
the higher one. Therefore, as far as we follow the SV reading, the tone of the AV stage should be
the higher counterpart of tone 6 (nă ̣ng tone)―that is, tone 5 (sắc tone).6 The tone change from
high to low in this case is due to the voicedness of the presyllable *a-.

6 Three high/low pairs of the modern 6 tones are tone 1 (ngang) / 2 (huyền), tone 3 (hỏ i) / 4 (ngã), and
tone 5 (sắc) / 6 (nă ̣ng).
140 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015)

(10)
b. PVM AV 17C 20C NV
vội - *a-βo:j5 βo:j6 vo:j6 vo:j6
阿盃 *a1 ɓo:j1
The case in (11a) also has quite important problems concerning spirantization and tone
height. The word vỗ về in the present orthography is an example of reduplication, but each syl-
lable was disyllabic in AV.
(11) a. Ch: 寵弄 鎮7 能 歡
CN: {阿*普}{阿*批}8 没 養 咍 盃
vỗ về một dường hay vui
comfort once know pleased
‘comfort and feel pleased for a while’ (14a4)
The SV initial transcribing of major syllables 普 phổ /fo:3/ and 批 phê /fe:1/ comes from MC
*pʰ (滂母). As far as we follow the order of spirantization → voicing proposed by Ferlus (1982),
the voiceless feature of the initial ph shows the middle stage of the process *p > ɸ > β. Meanwhile,
the tones of SV phổ (3: hoỉ tone) and phê (1: ngang tone) belong to the higher series, while those
of vỗ về (4: ngã tone, 2: huyền tone) belong to the lower, and both are the counterparts of the
opposite in each pair: 3/4 and 1/2. These conditions allow us to reconstruct the value of those
Chữ Nôm at the AV stage as in (11b):
(11)
b. PVM AV 17C 20C NV
vỗ - *a-ɸo:3 (βo:4)9 vo:4 vo:4
阿普 *a1 fo:3
về - *a-ɸe:1 (βe:2) ve:2 ve:2
阿批 *a1 fe:1
This case will be evidence for reconstructing both ɸ and β at the stage of AV. The spirantiza-
tion process and the value of labials at the AV stage are as follows in (12):
(12)
PVM AV 17C 20C NV
*-p- > *-ɸ- > β- > v- > v-
*-b- > *-β-
So far, I have applied the method described in section 3.1 to reconstruct the value of the
voiceless fricative *s and the labial stops *p and *b of PVM that had gone through the process of
spirantization. It is noteworthy that the cases introduced above are quite characteristic in that all
of them are disyllabic words and the structure of their Chữ Nôm characters shows the disyllabic-
ity clearly by using two phonetic components for each word.

4.  Reconstruction of AV
Hereafter, we apply the method proposed in section 3.1 for all the initial consonants of PVM.

7 In other texts, 鎮 is substituted by 纔 ‘for a while’ (Makita 1976: 56).

8 In the case that Chữ Nôm fonts are not available, we separate the components into {A*B} for vertical and
{A+B} for horizontal placement: e.g., {山*石} = 岩, {石+少} = 砂.
9 Not found in the Dictionary of Rhodes.
 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015) 141

Before this, it is necessary to discuss the syllable structure of AV.

4.1  Syllable structure

Ferlus (2004) supposed the syllable structure of PVM to be as follows:
(13)
PVM a. disyllables: C1-C2V(C3)
b. monosyllables:  C1V(C2)
The reason C1-C2V(C3) is regarded as disyllabic is because C1- in (13a) is usually followed by
a shwa, which had no phonological value inserted between C1 and C2. As we saw in section 3.2,
the spirantization process itself had not been completed and was still in the course of change at
the stage of AV. Therefore, we believe there is strong evidence for confirming that the presyllables
and major syllables had formed one unit so close as to interact with each other. This opinion can
be summarized as follows in (14):
PVM   AV 20C
(14)
C1-C2V(C3) > C1-C2V(C3)/T > C1V/T + C2V(C3)/T10 > C2V(C3)/T

We can make the AV version of the syllable types as in (15):

(15)
AV a. disyllables: C1-C2V(C3)/T
b. monosyllables: C2V(C3)/T

4.2  Presyllables (C1)

In the present material, Chữ Nôm characters such as 可, 車, 舎, 破, 司, 阿, and 羅, etc., can be
regarded as transcribing the presyllables. First of all, the presyllable list of the PVM stage, shown
in (16), was proposed by Ferlus (2007):
(16)
PVM *p- *t- *c- *k- *ʔa-
*s-
(*m-) (*l-)
(*r-)
The phonemes in ( ) are limited in number.

The examples in (17) are of Chữ Nôm characters transcribing disyllabic words:
(17)
CN SV PVM 20C (North) Ch Pages
*ka-
可汝 khả nhữ /kʰa:3 ɲɨ:4/11 *k-ɲəːʔ nhớ /ɲɤ:5/ 憶 15b1
*ca-
車莽 xa man ̃ g /tʃa:1 ma:ŋ4/ *c-maŋʔ mắng /maŋ5/ 聞 5b3, ...
舎莽 xá mãng /tʃa:5 ma:ŋ4/ *c-maŋʔ mắng /maŋ5/ 諦 23a5

10 Some fossilized disyllabic words in modern Vietnamese are mồ hôi ‘sweat’ (Thavung: pahû:), chiêm bao
‘nightmare’ (Ruc: cəpoː¹), etc.
11 The IPA transcription in this study is generally based on the modern Northern speech, except here, the
SV sounds of each Chinese character are based on the reconstructed AV phonemes. This is to clarify that
the AV phonetic values of the presyllables and the reconstructed PVM sounds are quite close.
142 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015)

*pa-
破散 ́ /pʰa:5 sa:n5/ *p-səɲʔ rắn /zan5/ 蛇 29a5
phá tan
*ta-/sa-
盃司 tư bôi /tɨ:1 ɓo:j1/ *t-puːj vui /vu:j1/ 歡,楽 20a3, 46a3
{司*盃}
*a-
阿盃 a bôi /a:1 ɓo:j1/ vô ̣i /vo:j6/ 早 38a1
{阿*普} a phổ /a:1 ɸo:3/ vỗ /vo:4/ 寵弄 14a4
{阿*批} a phê /a:1 ɸe:1/ về /ve:2/ 寵弄 14a4
阿吟 a ngâm /a:1 ŋʌm1/ *-ŋamʔ (?) ngậm /ŋʌm6/ 甘 42a4
*la-
羅打 la đả /la:1 ɗa:3/ *l-taːʔ đá /ɗa:5/ 丘山 41b2
In summary, the list of presyllables (C1-) extracted from the present materials will be as
follows:
(18)
AV *p- *t- *c- *k- *ʔa-
*s-
(*l-)

4.3  Initials of major syllables (C2)

According to Ferlus (2009), the list of initials of major syllables (C2) and their modern reflected
forms are as follows in (19):
(19) a. aspirated stops, voiceless fricatives
pʰ tʰ s kʰ h
> ph > th > t~r > kh >h
b. unaspirated stops p b t d c ɟ tʃ k g ʔ
> b~v > đ~d > ch~gi > x~gi > c~g >#
c. implosives ɓ ɗ ʄ
> m > n > nh
d. nasals m n ɲ ŋ
> m > n > nh > ng/ngh
e. voiced fricatives v j
> v > d
f. liquids r l
> r > l
In the case of the PVM *s, for instance, the modern reflection on the left t is the phoneme in
the monosyllabic words, while the modern reflection on the right r is the phoneme after it has
gone through spirantization in the disyllabic words. Accordingly, in the other cases, such as *p/*b
> b~v, *t/*d > đ~d, *c/*ɟ > ch~gi, *tʃ > x~gi, and *k/*g > c/k/q~g/gh, the italic letters on the left are
those in monosyllabic words, and those on the right are in disyllabic words (Ferlus 1982).
Hereafter, the AV phonetic values of (19) a~f will be reconstructed in order.
 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015) 143

4.3.1  Aspirated stops and voiceless fricatives

[Link] *pʰ
The PVM aspirated voiceless bilabial stop *pʰ corresponds to the modern spelling ph, and in the
17th century, it remained an aspirated stop, not a fricative as in the modern Northern dialect [f ].
Its process of change can be illustrated as in (20):
(20)
PVM AV  17C 20C (North)
*pʰ- > *pʰ- > pʰ- > f- (ph)
The correspondence of Chữ Nôm readings (ph) and the SV readings of their phonetic com-
ponents (ph) are as follows in (21):
(21) (MC: *pʰ 滂, *f 非, *fʰ 敷 > ) ph : (PVM: *pʰ > ) ph
CN Mod. Orth. SV MC Ch Pages
派 phaỉ /fa:j3/ 派 phaí /fa:j5/ 滂 被 19b4
沛 phải /fa:j3/ 沛 phái /fa:j5/ 滂 若 7b3, ...
鋪 phô /fo:1/ 鋪 phô /fo:1/ 滂 我等 30b5
...
{火+非} phơi /fə:j1/ 非 phi /fi:1/ 非 曝 20a2
發 phút /fu:t5/ 發 phát /fa:t5/ 非 須臾 17b3
番 phen /fɛ:n1/ 番 phan /fa:n1/ 敷 回,度, ... 8b2,10b3,...
Another correspondence is as follows in (22):
(22) (*b並 > ) b : (*pʰ > ) ph
CN Mod. Orth. SV MC Ch Pages
裴 phẩy /fʌj3/ 裴 bùi /ɓu:j2/ 並 拂拭 35b3
哺 phô /fo:1/ 哺 bô ̣ /ɓo:6/ 並 汝等 11b2
Most of the examples show the correspondence with the MC initials *pʰ (滂母, 非母, 敷
母)―that is, (pʰ > ) f ( ph). However, only two cases have a correspondence with (*b 並 > ) b.
This should not be regarded as an exception but as evidence of the stop feature of the phoneme.

[Link]  *tʰ
The PVM *tʰ corresponds to the modern spelling th, and its value remained tʰ throughout its
history.
(23)
PVM AV 17C 20C (North)
*tʰ- > *tʰ- > tʰ- > tʰ- (th)
The correspondence of the Chữ Nôm readings (th) and the SV readings of their phonetic
components (th) are as follows in (24):
(24) (*dʑ 船, *ɕ 書, *ʑ 常, *ts 清, *tʰ 透 > ) th : (*tʰ > ) th
CN Mod. Orth. SV MC Ch Pages
舌 thi ̣t /tʰi:t8/ 舌 thiê ̣t / tʰiət8/ 船 肉 29b5, ...
世 thay /tʰaj1/ 世 thế /tʰe:5/ 書 無量 3a3, ...
審 thẳm /tʰam3/ 審 thẩm /tʰʌm3/ 書 深 16b1
...
144 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015)

蜍 thờ /tʰɤ:2/ 蜍 thừ /tʰɨ:2/ 常 事 42b4, ...

沮 thả /tʰa:3/ 且 thả /tʰa:3/ 清 行 17a1
崔 thôi /tʰo:j1/ 崔 thôi /tʰo:j1/ 清 捨 33b1
体 thấy /tʰʌj5/ 体 thể /tʰe:3/ 透 見 4b1, ...
他 tha /tʰa:1/ 他 tha /tʰa:1/ 透 抜 23a3, ...
...
In all cases, the phoneme corresponds to the SV tʰ (th) from the MC initials 船, 書, 常, 清
and 透 and remains with the value tʰ. From this, it is likely that the SV readings of the initials 船,
書, 常, and 清 had also already become stops, not fricatives or affricates.

[Link] *s
The PVM *s had gone through the process of spirantization, so *s > r occurred in disyllabic words
(4c), and *s > t occurred in monosyllabic words (5c). As we have already seen in (6a) and (6b), the
value *s possibly remained in disyllabic words. First of all, other examples that went through the
process *s > r in the disyllabic words include those shown in (25) to (29):
(25) (*s 心, *ts 精 > ) t : (*s > ) r
CN Mod. Orth. SV MC Ch Pages
{塞+个} ráy /zaj5/ 塞 tái /ta:j5/ 心 洗濯 15a1, ...
{文(=塞)+个} ráy /zaj5/ 塞 tái /ta:j5/ 心 洗濯 11b5
{氵+塞} raý /zaj5/ 塞 taí /ta:j5/ 心 洗濯 17b5
索 rać h /zɛ:k5/ 索 tać /ta:k5/ 心 裂 29b5
燥 ráo /za:w5/ 燥 táo /ta:w5/ 心 乾 14a1, ...
{藻-氵} ráo /za:w5/ 藻 táo /ta:w5/ 精 - 29b5
(26) (*s 心, *z 邪 > ) t : (*s > ) t
CN Mod. Orth. SV MC Ch Pages
些 ta /ta:1/ 些 ta /ta:1/ 心 我 5a3, ...
細 tới /tɤ:j5/ 細 tế /te:5/ 心 向 15b2, ...
...
席 tiếc /tiək7/ 席 tích /ti:k7/ 邪 - 36a2
辞 từ /tɨ:2/ 辞 từ /tɨ:2/ 邪 慈 13b5
...
(27) (*s 心 > ) s : (*s > ) t
CN Mod. Orth. SV MC Ch Pages
珊 tan /ta:n1/ 珊 san /sa:n1/ 心 消散 9a4, ...
(28) (*ts 精, *dz 従 > ) t : (*s > ) t
CN Mod. Orth. SV MC Ch Pages
卒 tốt /to:t5/ 卒 tốt /to:t5/ 精 妙 2a3
斉 tầy /tʌj2/ 斉 tề /te:2/ 従 - 2a3
(29) (*t 端 > ) đ : (*s > ) t
CN Mod. Orth. SV MC Ch Pages
多 tơ /tɤ:1/ 多 đa /ɗa:1/ 端 帛 36b3
 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015) 145

In the case of monosyllabic words, the existence of the cases in (27) and (28) allow us to
reconstruct the value as *s―apart from one case where it is transcribed with the phoneme đ (29).
The process can be summarized as follows in (30):
(30)
PVM AV  17C 20C (North)
*-s- > *-s- > ɽ- > z- (r)
*s- > *s- > t- > t- (t)

[Link] *kʰ
The PVM *kʰ corresponds to the modern kh in orthography, but after the 17th century, it went
through the process kʰ > x (Gregerson 1969), illustrated in (31) below.
PVM AV 17C 20C (North)
(31)
*kʰ- > *kʰ- > kʰ- > x- (kh)
The correspondence of the Chữ Nôm readings (kh) and the SV readings of their phonetic
components (kh) are as follows in (32):
(32) (*kh 渓, *k 見 > ) kh : (*kʰ > ) kh
CN Mod. Orth. SV MC Ch Pages
看 khăn /xan1/ 看 khán /xa:n5/ 渓 辛苦 18a1
庫 khó /xɔ:5/ 庫 khố /xo:5/ 渓 苦 6a1, ...
...
{昌+圭} khoẻ /xwɛ:3/ 圭 khuê /xwe:1/ 見 健 13a4, ...
We can see only one case of correspondence with the MC *k (見母), but at the time of
SV formation, it already had the initial kh, so it can be treated the same way as the other cases.
Alternatively, it could be analogous to the SV reading of 奎, which also has kh as an initial.
Therefore, this phoneme also remained a stop until it became the fricative [x] after the 17th
century.

[Link] *h
The PVM *h corresponds to the modern h in orthography, and its value has remained h from
PVM until now, as shown in (33).
(33)
PVM AV  17C 20C (North)
*h- > *h- >  h- > h- (h)
The correspondence of the Chữ Nôm readings (h) and the SV readings of their phonetic
components (h, #) are as follows in (34) and (35):
(34) (*h 暁, *ɦ 匣, *ø 云, *k 見 > ) h : (*h > ) h
CN Mod. Orth. SV MC Ch Pages
咍 hai /ha:j1/ 咍 hai /ha:j1/ 暁 二 6b4, ...
歇 hết /he:t7/ 歇 hiết /hiət7/ 暁 盡 2a5, ...
...
鵠 ho ̣c /hɔ:k8/ 鵠 hộc /ho:k8/ 匣 習学 18a3
紇 hô ̣t /ho:t8/ 紇 hô ̣t /ho:t8/ 匣 露珠 9a3, ...
146 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015)

...
矣 hỡi /hɤ:j4/ 矣 hĩ /hi:4/ 云 嗚呼 34a5
忽 hót /hɔ:t7/ 忽 hốt /ho:t7/ 見 除 15a3
(35) (*Ɂ 影 > ) # : (*h > ) h
CN Mod. Orth. SV MC Ch Pages
謁 hết /he:t7/ 謁 yết /iət7/ 影 終 18a4, ...
In most cases, the phoneme corresponds to the SV readings for the MC initials *h (暁母)
and *ɦ (匣母). In only one case each does it correspond to *ø (云母), *k (見母), and *Ɂ (影母).
The SV readings for the former two MC initials are always h in the modern orthography. The
only exception―MC initial *Ɂ (影母)―could also be analogous to 歇 hết as it shares the same
phonetic component. Therefore, the value h has remained quite steady from PVM until today.
To summarize all the reconstructed forms of aspirated stops and voiceless fricatives, we have
the chart in (36):
(36) aspirated stops and voiceless fricatives
PVM AV 17C 20C (North)
*pʰ- *pʰ- > pʰ > f (ph)
*tʰ- *tʰ- > tʰ > tʰ (th)
*s- *-s- > ɽ > z (r)
*s- > t > t (t)
*kʰ- *kʰ- > kʰ > x (kh)
*h- *h- > h > h (h)

4.3.2  Unaspirated stops

The series of unaspirated stops went through the process of spirantization, except for the case
of the glottal stop. These processes were shown in (5) and (6). Below are the actual cases of Chữ
Nôm transcription.

[Link]  *p and *b
For the case of the bilabial stops *p and *b, we have shown the process of disyllables in (12). As
far as we can attest the trace of voiceless fricatives shown in (11), we can assume that the spiran-
tization process was still under way at this point.
The following in (37) are the other cases of the labials, with the cases of the monosyllables
that have nothing to do with the process of spirantization listed first.
(37) (*p 幫, *b 並 > ) b : (*p, *b > ) b
CN Mod. Orth. SV MC Ch Pages
巴 ba /ɓa:1/ 巴 ba /ɓa:1/ 幫 三 5b1, ...
奔 buôn /ɓuən / 奔 bôn /ɓo:n / 幫
1 1
經商 19b1
...
盆 buồn /ɓuən2/ 盆 bồn /ɓo:n2/ 並 懊惱 38a2
倍 bụi /ɓu:j6/ 倍 bội /ɓo:j6/ 並 塵 35b3, ...
...
 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015) 147

In the case of monosyllabic words, most of the cases transcribe the initial with b corre-
sponding to the MC initials *p (幫母) and *b (並母). However, we have no evidence to confirm
whether it had already been voiced or still remained voiceless. In general, as far as the Chữ Nôm
characters were created using the SV readings that distinguish voiced and voiceless consonants
by the height of tones―i.e., higher tones and lower ones―one can be sure that the bipartition of
high and low tones had already been completed. It follows that the contrast between voiced and
voiceless had already disappeared, at least within Vietnamese. Here we continue with the suppo-
sition that the phoneme in question had already been voiced.
The cases in (38) and (39) are those of the disyllables that had gone through the process of
spirantization.
(38) (*p 幫, *b 並 > ) b : (*p, *b > ) v
CN Mod. Orth. SV MC Ch Pages
邦 vâng /vʌŋ1/ 邦 bang /ɓa:ŋ1/ 幫 奉 6b2, ...
包 vào /va:w2/ 包 bao /ɓa:w1/ 幫 入 5a1, ...
...
{女+夂*田} vợ /vɤ:6/ 備 bi ̣ /ɓi:6/ 並 婚娵 19b2
卞 vin /vi:n1/ 卞 biê ̣n /ɓiən6/ 並 攀 11a2
(39) (*pʰ 滂 > ) b : (*p, *b > ) v
CN Mod. Orth. SV MC Ch Pages
丕 vậy /vʌj6/ 丕 phi /fi:1/ 滂 故 2a4, ...
Again, the case in (39) shows the stage when the voicing of the voiceless fricatives had not
occurred.
In (40), we summarize the above discussion.
(40)
PVM AV 17C 20C (North)
*-p- > *-ɸ- β- > v- (v)
*-b- > *-β-
*p- *ɓ- > ɓ- > ɓ- (b)
*b-

[Link]  *t and *d
The PVM dental stops also went through spirantization in the disyllabic words. Sections [Link]
to [Link] present cases of monosyllabic words.
(41) (*t 端, *d 定 > ) đ : (*t, *d > ) đ
CN Mod. Orth. SV MC Ch Pages
怛 đất /ɗʌt7/ 怛 đát /ɗa:t7/ 端 土 3b1, ...
帯 đái /ɗa:j5/ 帯 đái /ɗa:j5/ 端 不浄 11b5, ...
...
畑 đen ̀ /ɗɛ:n2/ 田 điền /ɗiən2/ 定 燈 26b2
調 đều /ɗe:w2/ 調 điều /ɗieu2/ 定 惣 6a1, ...
...
For the same reason as with the labials, we suppose the contrast between voiced and voiceless
148 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015)

disappeared and merged into a voiced stop.

The cases that went through the spirantization process are listed in (42).
(42) (*t 端, *d 定 > ) đ : (*t, *d > ) d
CN Mod. Orth. SV MC Ch Pages
{心+多} da /za:1/ 多 đa /ɗa:1/ 端 質 15a1
帯 dưới /zɨəj5/ 帯 đái /ɗa:j5/ 端 下 10b1, ...
...
代 dạy /zaj6/ 代 đại /ɗa:j6/ 定 訓教 18a3, ...
杜 dỗ /zo:4/ 杜 đỗ /ɗo:4/ 定 訓教 18a3
...
The cases of consonant d were here transcribed with the SV reading đ corresponding to the
MC initials *t (端母) and *d (定母) without exception. Therefore, it is different from the case of
labials in that there are no cases transcribing the voiceless fricatives. The process is as follows in
(43):
PVM AV 17C 20C (North)
(43)
*-t- > - ð- > z- (d)
*-d- > *-ð-
*t- *ɗ- > ɗ- > ɗ- (đ)
*d-

[Link]  *c and *ɟ
The PVM palatal stops *c and *ɟ also went through the process of spirantization. The cases in
(44) and (45) are those of monosyllable words that kept the stop feature.
(44) (*tɕ 章 > ) ch : (*c, *ɟ > ) ch
CN Mod. Orth. SV MC Ch Pages
沚 chảy /tʃaj3/ 止 chỉ /tʃi:3/ 章 流 8b2, ...
眞 chân /tʃʌn1/ 眞 chân /tʃʌn1/ 章 脚 11a3, ...
...
(45) (*ʈ 知, *ɖ 澄, *tʂ 荘 > ) tr : (*c, *ɟ > ) ch
CN Mod. Orth. SV MC Ch Pages
吒 cha /tʃa:1/ 吒 trá /tʃa:5/ 知 父 5a1, ...
徴 chưng /tʃɨ:ŋ1/ 徴 trưng /tʃɨ:ŋ1/ 知 爲 5a2, ...
召 chi ̣u /tʃi:w6/ 召 triê ̣u /tʃiəw6/ 澄 受 11b3, ...
{宀*丁} chữ /tʃɨ:4/ {宀*丁} trữ /tʃɨ:4/ 澄 文 5a5, ...
庄 chẳng /tʃaŋ3/ 庄 trang /tʃa:ŋ1/ 荘 不 5b4, ...
債 cháy /tʃaj5/ 債 trái /tʃa:j5/ 荘 火 16a2, ...
The SV reflection of the MC initial *tɕ (章母)―that is, ch―along with the SV initials with
nearby places of articulation such as tr (reflection for the MC initials *ʈ 知, *ɖ 澄, and *tʂ 荘)
make it clear that the PVM *c and *ɟ had merged into voiceless stops. The cases of disyllabic
words are shown in (46) and (47).
 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015) 149

(46) (*tɕ 章 > ) ch : (*c, *ɟ > ) gi

CN Mod. Orth. SV MC Ch Pages
折 giết /ziət5/ 折 chiết /tʃiət5/ 章 屠 13a3, ...
(47) (*ɖ 澄 > ) tr : (*c, *ɟ > ) gi
CN Mod. Orth. SV MC Ch Pages
陣 giận /zʌn6/ 陣 trận /tʃʌn6/ 澄 嗔恨 19a4
{宀*丁} giữ /zɨ:4/ {宀*丁} trữ /tʃɨ:4/ 澄 守護 11b2, ...
{木+長} giường /zɨəŋ2/ 長 trường /tʃɨəŋ2/ 澄 床 21a1
...
As far as this material is concerned, all the consonants in question are transcribed with ch and
tr. Therefore, the process can be traced as follows in (48):
PVM AV MV 20C (North)
(48)
*-c- > *-ç- ȷ- > dʒ- (gi)
*-ɟ- >   -
*c- *c- > c- > c- (ch)
*ɟ-

[Link] *tʃ
The PVM affricate *tʃ had become x in all cases. According to Ferlus (2009), it became gi in
disyllabic words. The cases of monosyllables are listed in (49) to (52).
(49) (*tɕʰ 昌 > ) x : (*tʃ > ) x
CN Mod. Orth. SV MC Ch Pages
車 xa /sa:1/ 車 xa /sa:1/ 昌 遠 16b2, ...
侈 xé /sɛ:5/ 侈 xỉ /si:3/ 昌 研 29b3
...
(50) (*tɕ 章 > ) ch : (*tʃ > ) x
CN Mod. Orth. SV MC Ch Pages
{示+占} xem /sɛ:m1/ 占 chiêm /tʃiəm1/ 章 觀 17a1, ...
嗔 xin /si:n1/ 眞 chân /tʃʌn1/ 章 須 33b5, ...
(51) (*ɕ 書, tsʰ 清 > ) th : (*tʃ > ) x
CN Mod. Orth. SV MC Ch Pages
拭 xức /sɨ:k7/ 拭 thức /thɨ:k7/ 書 塗 8a1, ...
千 xiên /siən1/ 千 thiên /thiən1/ 清 刺 29b3
崔 xui /su:j1/ 崔 thôi /tho:j1/ 清 被 19a5
(52) (ʈʰ 徹 > ) s : (*tʃ > ) x
CN Mod. Orth. SV MC Ch Pages
撑 xanh /sɛ:ŋ1/ 撑 sanh /sɛ:ŋ1/ 徹 - 15a2
丑 xấu /sʌw5/ 丑 sửu /sɨ:w3/ 徹 醜状 21b3
The existence of the consonant ch for the MC initial *tɕ (章母) to transcribe x in (50) shows
the stage of *tʃ, while those cases in (52) show the value of ʃ in the process *tʃ > ʃ. Concerning the
150 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015)

consonant x in the SV readings, the general reflection of the MC initial *ɕ (書母) is th, but when
it appears with medial j, the SV reflection is x. Meanwhile, the value of x in the 17th century
was ʃ, and the value of th in modern Vietnamese is so “soft” that it is sometimes pronounced θ.
Therefore, when it comes with medial j, the value can be quite close to that of x.
Unfortunately, we did not find any cases that had gone through the spirantization process in
the present material, as depicted in (53).
(53)
PVM AV   17C 20C (North)
*-tʃ >  - > ȷ- > dʒ- (gi)
*tʃ > *tʃ/ʃ > ʃ- > s- (x)

[Link]  *k and *g
In the case of velars, the SV reflection of the MC *k (見母) and *ɡ (群母) is c/k/q, not g/gh.
However, when the SV readings are further Vietnamized in the colloquial language, it often
becomes g/gh.
The cases of the monosyllable words are presented in (54).
(54) (*k 見, *ɡ 群 > ) c/k/q : (*k, *ɡ > ) c/k/q
CN Mod. Orth. SV MC Ch Pages
核 cây /kʌj1/ 該 cai /ka:j1/ 見 樹 29b2
拮 cất /kʌt7/ 吉 cát /ka:t7/ 見 挙 22b3
…
其 caỳ /kaj2/ 其 kì /ki:2/ 群 犂 31a3
揆 cỡi /kɤ:j4/ 揆 quĩ /kwi:4/ 群 辺 44b3
…
As in the case of the SV readings, the PVM labials and dentals became voiced implosives,
while velars became voiceless c/k/q. This is one of the typical features in Vietnamese.
The cases of the disyllabic words are shown in (55) and (56).
(55) (*k 見, *ɡ 群 > ) c/k/q : (*k, *ɡ > ) g/gh
CN Mod. Orth. SV MC Ch Pages
斤 gần /ɣʌn2/ 斤 cân /kʌn1/ 見 近 16b3
丐 gái /ɣa:j5/ 丐 cái /ka:j5/ 見 女 7b4, ...
…
及 gặp /ɣap8/ 及 cập /kʌp8/ 群 遭値 3a3
技 ghẽ /ɣɛ:4/ 技 kĩ /ki:4/ 群 違 31a4, ...
…
(56) (*kʰ 溪 > ) kh : (*k, *ɡ > ) g/gh
CN Mod. Orth. SV MC Ch Pages
可 gả /ɣa3/ 可 khả /xa:3/ 渓 婿 22a2
{子+可} gã /ɣa4/ 可 khả /xa:3/ 渓 孩子 13b5
In the 17th century, for instance, the velar consonants retained the contrast between voiced
and voiceless (k/g), and the voiced consonant had become a fricative as in the modern language.
Here, it is quite clear from the modern reflection that the velar stops remained voiceless in the
 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015) 151

monosyllabic words, while those in the disyllabic words became voiced. Here again, the existence
of the cases in (56) show the medial stage in the process *k > x > ɣ, which is illustrated in (57).
(57)
PVM AV MV 20C (North)
*-k- > *-x- ɣ- > ɣ- (g/gh)
*-ɡ- > *-ɣ-

*k- *k- > k- > k- (c/k/q)

*ɡ-

[Link] *ʔ
The glottal stops are never transcribed in Chữ Nôm or modern orthography. They are transcribed
with the SV readings of the MC initial *ʔ (影母), as demonstrated in (58).
(58) (*ʔ 影 > ) # : (*ʔ > ) #
CN Mod. Orth. SV MC Ch Pages
咹 ăn /an1/ 安 an /a:n1/ 影 食 10a5, ...
意 ấy /ʌj5/ 意 ý /i:5/ 影 爾 23a3, ...
 ...
Now let us summarize the process of the stop series in PVM as shown in (59).
(59)
PVM AV 17C 20C (North)
*p- *ɓ- > ɓ- > ɓ- (b)
*b-
*-p- > *-ɸ- β- > v- (v)
*-b- > *-β-
*t- *ɗ- > ɗ- > ɗ- (đ)
*d-
*-t- > - ð- > z- (d)
*-d- > *-ð-
*c- *c- > c- > c- (ch)
*ɟ-
*-c- > *-ç- ȷ- > dʒ- (gi)
*-ɟ- > -
*tʃ- > *tʃ-/ʃ- > ʃ- > s- (x)
*-tʃ- > - > ȷ- > dʒ- (gi)
*k- *k- > k- > k- (c/k/q)
*ɡ-
*-k- > *-x- ɣ- > ɣ- (g/gh)
*-ɡ- > *-ɣ-
*#- > *#- > #- > #- (ø)

4.3.3 Implosives
[Link] *ɓ
Some part of the phonemes transcribed with m in the modern orthography can be traced back to
152 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015)

the PVM implosive *ɓ. By referring to the word list in Ferlus (2007), we can identify which ones
are which and find the same words in our material. The comparison is presented in (60).
(60) (*m 明 > ) m : (*ɓ > ) m
CN Mod. Orth. SV MC Ch Pages
黙 mắc /mak7/ 黙 mặc /mak8/ 明 掛 22b1
末 mất /mʌt7/ 末 mạt /ma:t8/ 明 損 11a1, ...
美 mở /mɤ:3/ 美 mĩ /mi:4/ 明 發 5b5, ...

At the stage of forming the SV readings, the PVM *ɓ had already become a nasal m.
Therefore, the value in the present material is also a nasal, as shown in (61).
PVM AV 17C 20C (North)
(61)
*ɓ- > *m- > m- > m- (m)

[Link] *ɗ
Using the same method as with the labials, we can find the words in question in our material.
(62) (*n 泥 > ) n : (*ɗ > ) n
CN Mod. Orth. SV MC Ch Pages
南 năm /nam1/ 南 nam /na:m1/ 泥 五 5b3, ...
悩 nấu /nʌw5/ 悩 não /na:w4/ 泥 烹 16a5, ...
餒 nuôi /nuəj1/ 餒 nỗi /no:j4/ 泥 養育 16a2, ...
All the cases in (62) are transcribed with nasal n; therefore, the process can be summarized as
in (63).
(63)
PVM AV 17C 20C (North)
*ɗ- > *n- > n- > n- (n)

[Link] *ʄ
Another case is the palatal implosive. We can find only one word of this case. The word and cor-
responding rule are shown in (64) and (65), respectively.
(64) (*ɲ 日 > ) nh : (*ʄ > ) nh
CN Mod. Orth. SV MC Ch Pages
饒 nhau /ɲaw1/ 饒 nhiêu /ɲiəw1/ 日 諸 5a2, ...
(65)
PVM AV 17C 20C (North)
*ʄ- > *ɲ- > ɲ- > ɲ- (nh)
Let us summarize the process of implosives as follows in (66):
(66)
PVM AV 17C 20C (North)
*ɓ- > *m- > m- > m- (m)
*ɗ- > *n- > n- > n- (n)
*ʄ- > *ɲ- > ɲ- > ɲ- (nh)
 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015) 153

4.3.4 Nasals
The next series are those of the PVM nasals. Essentially, all these phonemes had not changed
their values but, concerning the labials, dentals, and palatals, had merged with the implosive
series at the AV stage. Examples are presented for each phoneme in the sections [Link] to
[Link].

[Link] *m
(67) (*m 明 > ) m : (*m > ) m
CN Mod. Orth. SV MC Ch Pages
沫 mắt /mat7/ 沫 mạt /ma:t8/ 明 涙 12b4(nước+), ...
卯 máu /maw5/ 卯 mão /ma:w4/ 明 血 8b2, ...
{女+美} me ̣/mɛ:6/ 美 mĩ /mi:4/ 明 母 4b5, ...
...

[Link] *n
(68) (*n 泥 > ) n : (*n > ) n
CN Mod. Orth. SV MC Ch Pages
尼 nay /naj2/ 尼 ni /ni:1/ 泥 今 2a4, ...
喃 nắm /nam5/ 喃 nam /na:m1/ 泥 胞 9b2
...

[Link] *ɲ
(69) (*ɲ 日 > ) nh : (*ɲ > ) nh
CN Mod. Orth. SV MC Ch Pages
如 nhơ /ɲɤ:1/ 如 như /ɲɨ:1/ 日 穢 15a4
茹 nhà /ɲa:2/ 茹 như /ɲɨ:1/ 日 宅 16a2, ...
二 nhe ̣ /ɲɛ:6/ 二 nhi ̣ /ɲi:6/ 日 輕 7b4, ...
...
(70) (*j 以 > ) d : (*ɲ > ) nh
CN Mod. Orth. SV MC Ch Pages
野 nhả /ɲa:3/ 野 dã /za:4/ 以 吐 11b4
{口+也} nhả /ɲa:3/ 也 dã /za:4/ 以 吐 13b1, ...
耶 nhã /ɲa:4/ 耶 da /za:1/ 以 毀辱 18b3(nhục+)
As for the PVM *ɲ, it is not only the SV nh for the MC *ɲ (日母) but also the SV d for the
MC *j (以母) that corresponds to this phoneme. This is exactly the same as the case of nhiếc,
which is alternated by diếc in the modern language, and in the Dictionary of Rhodes, where nhà
has an alternative transcription of da,̀ etc. Therefore, the cases in (70) show the same kind of
alternation.

[Link] *ŋ
(71) (*ŋ 疑 > ) ng/ngh : (*ŋ > ) ng/ngh
CN Mod. Orth. SV MC Ch Pages
{女+疑} ngờ /ŋɤ:2/ 疑 nghi /ŋi:1/ 疑 惑 41a4
154 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015)

我 ngã /ŋa:4/ 我 ngã /ŋa:4/ 疑 倒 46a1

外 ngồi /ŋo:j2/ 外 ngoại /ŋwa:j6/ 疑 坐 17b1, ...
...
The process of nasal series is as follows in (72):
(72)
PVM AV   17C 20C (North)
*m- > *m- > m- > m- (m)
*n- > *n- > n- > n- (n)
*ɲ- > *ɲ- > ɲ- > ɲ- (nh)
*ŋ- > *ŋ- > ŋ- > ŋ- (ng/ngh)

4.3.5  Voiced fricatives

[Link] *v
A part of the modern v goes back to the PVM *v. Here again, we refer to Ferlus (2007) to select
the samples in our material, shown in (73).
(73) (*ø 云, *ɱ 微 > ) v : (*v > ) v
CN Mod. Orth. SV MC Ch Pages
衛 về /ve:2/ 衛 vê ̣ /ve:6/ 云 歸 7a3, ...
{尾+个} vẫy /vʌj4/ 尾 vı̃ /vi:4/ 微 攪 11a4
It is noteworthy that when the PVM *p and *b went through spirantization and became v,
they are transcribed with bilabial b, while the PVM *v is transcribed with fricative v. Therefore,
Chữ Nôm transcription distinguishes clearly between the PVM bilabials and labial fricatives; see
(74).
(74)
PVM AV   17C 20C (North)
*v- > *v- > v- > v- (v)

[Link] *j
A part of modern d is traced back to the PVM *j, as demonstrated in (75) and (76).
(75) (*j 以 > ) d : (*j > ) d
CN Mod. Orth. SV MC Ch Pages
酉 dâu /zʌw1/ 酉 dậu /zʌw6/ 以 妾 21b4
{曳+｀} dậy /zʌj6/ 曳 duê ̣ /zwe:6/ 以 起 17b1, ...
{曳+｀} dài /za:j2/ 曳 duê ̣ /zwe:6/ 以 長 31a3, ...
(76)
PVM AV   17C 20C (North)
*j- > *j- > ᶁ- > z- (d)
The process of voiced fricatives is as follows in (77):
PVM AV   17C 20C (North)
(77)
*v- > *v- > v- > v- (v)
*j- > *j- > ᶁ- > z- (d)
 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015) 155

4.3.6 Liquids
[Link] *r
A part of the modern r is also traced back to the PVM *r, as demonstrated in (78).
(78) (*l 來 > ) l : (*r > ) r
CN Mod. Orth. SV MC Ch Pages
{肉+聿} ruột /zuət8/ 律 luật /lwʌt8/ 來 腸 22b1, ...
{酉+留} rượu /zɨəw6/ 溜 lựu /lɨw6/ 來 酒 20b3

[Link] *l
The value of the PVM *l had not changed throughout its history, as shown in (79).
(79) (*l 來 > ) l : (*l > ) l
CN Mod. Orth. SV MC Ch Pages
凌 lưng /lɨ:ŋ1/ 凌 lăng /laŋ1/ 來 身 24b2
礼 lấy /lʌj5/ 礼 lễ /le:4/ 來 以 22a3, ...
羅 là /la:2/ 羅 la /la:1/ 來 是 7a3, ...
...
The process of liquids is as follows in (80):
(80)
PVM AV  17C 20C (North)
*r- > *r- > r- > r- (r)
*l- > *l- > l- > l- (l )

4.4  Consonant clusters

Another topic concerning the consonant system of AV is the pattern of consonant clusters.
Ferlus (1992) proposed the list of patterns of consonant clusters shown in (81):
(81)
PVM pr br tr dr kr gr
pl bl ml kl gl
kj
In our material, the cases in (82) offer quite clear evidence for how the patterns of consonant
clusters should be reconstructed.
(82)
CN SV MV 20C   Ch Pages
bl- > ʈ- (tr)
巴邏 ba la /ɓa:1 la:1/ blả trả 報答 8b5
波來 ba lai /ɓa:1 la:j1/ blái trái - 11a2-3
婆論 bà luận /ɓa:2 lwʌn6/ blọn trọn 全備 11b5
pʰl- > ʈ- (tr)
坡栗 pha lật /fa:1 lʌt8/ blạt trật 失 13b1
kʰr- > ʂ- (s)
可羅 khả la /xa:3 la:1/ sa sa 垂 28a1
156 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015)

可列 khả liệt /xa:3 liət8/ sắt sắt 鐵 29a5, 31a3

kr- > ʂ- (s)
{古*弄} cổ lộng /ko:3 lo:ŋ6/ sống sống 生, 存 6a3, 15b1, 30a1, 44a3
/ 古弄
呂巨 cự lữ /kɨ:6 lɨ:4/ sữa sữa 乳 34a4, 35a4
夌巨 cự lăng /kɨ:6 laŋ1/ sưng sưng 腫 38a2
pʰr- > ʂ- (s)
坡律 pha luật /fa:1 lwʌt8/ sốt sốt 熱 29a4
破律 phá luật /fa:3 lwʌt8/ sốt sốt 熾 29a3
br - > ʂ- (s)
呂巴 ba lữ /ɓa:1 lɨ:4/ sữa sữa 白血 18a2
ml- > l- (l)
麻例 ma lệ /ma:1 le:6/ mlời lời 言，教，語 19a1, 31a2, 37b4-5
麻隣 ma lân /ma:1 lʌn1/ lăn lăn 荏苒 19b2
麻碌 ma lộc /ma:1 lo:k8/ - lóc 荏苒 19b2
The extracted cluster patterns are as follows in (83):
(83)
PVM AV 17C 20C (North/South)
Cr- > *Cr- (pʰr, br, kr-kʰr) > ʂ- > s/ʂ- (s)
Cl- > *Cl- (pʰl, bl, kl) > bl- > tʃ/ʈ- (tr)
  (ml) > ml/l- > ɲ- (nh)/l- (l )

5. Conclusion
Through the analysis of the Chữ Nôm characters contained in the the Buddhist sutra Phâ ̣t Thuyết
Đa ̣i Báo Phụ Mẫu Ân Trọng Kinh 佛説大報父母恩重經, we reconstructed the phonemes of ini-
tial consonants and the syllable structure of Ancient Vietnamese (15C). Concerning the process
of spirantization, which can be attested quite clearly in the material, the phonological features of
the Chữ Nôm characters can be placed between the stage of sprirantization and that of voicing
for the medial consonants.
The results of the reconstruction are summarized as follows:
Syllable Structure
disyllables: C1-C2V(C3)/T
monosyllables:  C2V(C3)/T
Presyllables (C1)
p- t- c- k- ʔa-
s-
(l-)
Consonant Clusters
pʰr br kr (kʰr)
pʰl bl kl ml
 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015) 157

Initial Consonants (C2)

ɓ ɗ
c k ʔ
pʰ tʰ kʰ
β/ɸ v ð s ʃ ç j ɣ/x h
m n ɲ ŋ
r
l

References
Ferlus, Michel (1982) Spirantisation des obstruantes médiales et formation du système consonantique du
vietnamien. Cahiers de Linguistique Asie Orientale 11(1): 83–106.
Ferlus, Michel (1992) Histoire abrégée de l’évolution des consonnes initiales du Vietnamien et du
Sino-Vietnamien. Mon-Khmer Studies 20: 111–125.
Ferlus, Michel (2004) The origin of tones in Viet-Muong. In: Somsonge Burusphat (ed.) Papers from the
Eleventh Annual Meeting of the Southeast Asian Linguistics Society 2001, 297–313. Arizona State Univer-
sity.
Ferlus, Michel (2007) Lexique de racines Proto Viet-Muong [Proto Vietic Lexicon]. Unpublished manuscript.
([Link]
Ferlus, Michel (2009) A layer of Dongsonian vocabulary in Vietnamese. Journal of the Southeast Asian Lin-
guistics Society 1: 95–109.
Gregerson, Kenneth J. (1969) A study of Middle Vietnamese. BSEI 44: 135–193.
Hoàng, Thị Ngọ (1999) Chữ Nôm và tiếng Việt qua bản giải âm Phật thuyết Đại báo phụ mẫu ân trọng kinh.
Hà Nội: Nxb Khoa học Xã hội.
Makita, Tairyo (1976) A study on Pseudo-Sutras [疑經研究]. Kyoto: Kyoto Daigaku Jinbun Kagaku Ken-
kyujo.
Mineya, Toru (1972) A study of Sino-Vietnamese readings [越南漢字音の研究]. Toyobunko Ronso 53.
Tokyo: Toyobunko.
Mori, Hiromichi (1991) Ancient phonology and the creation of Nihonshoki [古代の音韻と日本書紀の成立].
Tokyo: Taishukan Shoten.
Nguyễn, Tài Cẩn (1995) Giáo trình lịch sử ngữ âm tiếng Việt. Hà Nội: Nxb. Giáo dục.
Nguyễn, Tài Cẩn (2008) Một số vấn đề về ngành Nôm học. Hội nghị Nôm học, 11–12 tháng 4 năm 2008.
Trung tâm Triết học, Văn hoá & Xã hội Việt Nam. Đại học Temple.
Shimizu, Masaaki (1996) On the Chu Nom characters contained in the Sino-Vietnamese text of Phat
Tuyet Dai Bao Phu Mau An Trong Kinh [漢文＝字喃文対訳『佛説大報父母恩重經』 に見る字喃につ
いて]. Ningen-Kankyogaku 5: 83–104. Kyoto University.
Shimizu, Masaaki (1999) Sino-Vietnamese readings included in the Dictionary of Alexandre de Rhodes
[Alexandre de Rhodesの辞書に見るベトナム漢字音について]. Tonanajia—Rekishi to Bunka 28: 55–80.
Japan Society for Southeast Asian Studies.
Trần, Tro ̣ng Dương (2011) Phâ ̣t Thuyết có phả i là dich
̣ phẩ m Nôm củ a thế kỷ XII? Ngôn ngữ 4 (263):
31–47.
Trần, Văn Kiê ̣m, L.m. An-tôn (2004) Giúp đo ̣c Nôm và Hán Viê ̣t. Nxb. Đà Nẵng. Chữ Nôm. Hoa Kỳ: Hội
Bả o tổ n Di sả n.
158 SHIMIZU Masaaki / NINJAL Research Papers 9: 135–158 (2015)

字喃資料を用いた上古ベトナム語頭子音の再構

清水政明
大阪大学／国立国語研究所 共同研究員

要旨
従来上古ベトナム語（AV）の再建にしばしば字喃が利用されてきたが，子音連続など限られた
音韻特徴が言及されるに止まった。本稿では，漢文・字喃文対訳『佛説大報父母恩重經』所収の
字喃を基礎データとし Ferlus（2009）の Proto Viet-Muong（PVM）の再構音を参照しつつ，AV 頭
子音の体系の再建を試みた。結果，2 音節語の分析により摩擦音化の過程が実証されたのみならず，
頭子音全体の体系を示すに至った。今後も同様の資料を分析することにより，PVM から中古ベ
トナム語（17C）への変遷を解明する重要な鍵が得られると期待される。

キーワード：上古ベトナム語，字喃，ベトナム漢字音